Discovery of a non-canonical prototype long-chain monoacylglycerol lipase through a structure-based endogenous reaction intermediate complex

Assigning physiological enzyme functions remains challenging despite the explosion of available protein sequences and structures. Many enzymes display broader or different substrate profiles in cells than in vitro, and structural studies typically require soluble ligands, limiting access to endogenous hydrophobic substrates. Lipases, which act at water–lipid interfaces and often employ a mobile lid/cap domain to control access to the active site, are important biocatalysts across biotechnology but their substrate specificity is difficult to predict due to lid diversity. An enzyme from the thermophilic anaerobe Thermoanaerobacter thermohydrosulfuricus (Tth) had been tentatively classified as a lipase with notable robustness, yet its molecular mechanism and true substrate scope were unknown. This study asks whether endogenous ligand capture during crystallization can reveal the physiological substrate and mechanism, enabling functional annotation and engineering of this orphan enzyme.

Background on lipases highlights their reversible hydrolysis/synthesis of glycerol esters, interfacial activation, and the role of lid domains in substrate access. Lid sequences are poorly conserved and structurally diverse, complicating specificity predictions. Extremophilic lipases are attractive for industry due to tolerance of harsh conditions. Prior studies often used soluble analogs or engineered solubilization; structures with authentic long-chain lipid ligands are rare. The Tth enzyme was previously suggested to be a lipase based on broad turnover and robustness but lacked mechanistic details. Comparative structural context includes α/β-hydrolase fold lipases and esterases, with some related enzymes (e.g., cinnamoyl and feruloyl esterases from Lactobacillus johnsonii and Butyrivibrio proteoclasticus) showing similar cores but different lid features and substrate specificities toward phenylpropanoids. Conventional MAG lipases (e.g., human MGL) generally lack the non-canonical minimal HBH lid architecture described here.

• Protein production: Wild-type and variant Tth MAG lipase expressed in E. coli (pETBlue-1 or pETM-14 vectors) including SeMet-labeled protein and lysine-methylated preparations; purification by Ni-NTA affinity and size-exclusion chromatography.
• Crystallography: Sitting-drop vapor diffusion at 20 °C; crystals obtained for native, PMSF-inhibited, SeMet, and methylated protein. Data collected at ESRF BM14 and EMBL/DESY beamlines. Phasing with SeMet SAD (SHELXD/E, PHENIX AutoSol), model building with ARP/WARP and COOT, refinement in PHENIX. Structures deposited (7Q4J, 7Q4H, 8B9S). AlphaFold2 (ColabFold) used for comparison.
• Endogenous ligand identification: Crystals washed, lipids extracted (chloroform/methanol), analyzed by LC–MS on C18 column; a single peak matched glycerol monostearate (m/z 359 in positive mode), consistent with a C18 MAG species.
• Biophysics: Size exclusion chromatography to assess oligomeric state; nanoDSF to determine fold melting temperature; circular dichroism to verify folding.
• Enzymatic assays:
  • MAG/DAG/TAG assay: Turnover of 1-octanoyl-rac-glycerol (C8 MAG), 1,2-dioctanoyl-sn-glycerol (C8 DAG), tricaprylin (C8 TAG), and 1-oleoyl-rac-glycerol (C18:1 MAG) at 40 °C; glycerol release quantified.
  • p-Nitrophenyl (pNP) acyl ester assay: Substrates C2–C18 in microplate format at 40 °C; initial rates at 410 nm with autohydrolysis correction. Michaelis–Menten kinetics across 0.005–1 mM to obtain KM, kcat, and kcat/KM.
  • In vivo activity: Expression in E. coli; extraction and methylation of fatty acids to FAMEs; analysis by HPTLC and GC–MS; semi-quantitative comparison across variants.
• Mutagenesis: Active site (S113A), glycerol-binding residue (E43A/K), dimer interface (E72R), lid deletion (Δ140–183), lid tunnel residue (Y154A/R); validation of folding, stability, and oligomeric state.
• Structural analysis: Mapping of lid tunnel, active site interactions (glycerol and MAG intermediate), comparison of native vs PMSF structures to assess lid conformations; comparison to related esterases and MAG lipases; analysis of conserved motifs (HGF, FSE/D).

• Structure/function discovery via endogenous ligand capture:
  • High-resolution structures revealed glycerol in the active site and an extended electron density consistent with a C18 MAG reaction intermediate traversing a 15 Å hydrophobic tunnel in a minimal α-helix–β-hairpin–α-helix (HBH) lid domain.
  • LC–MS of crystal extracts identified glycerol monostearate, supporting assignment of a C18 monoacylglycerol species.
  • Geometry and omit maps support a loosely coordinated tetrahedral MAG reaction intermediate near catalytic Ser113 and glycerol O1.
• Architecture and oligomerization:
  • The enzyme is a symmetric dimer (SEC and crystal), with a flat interface (~10% SASA) forming a cleft linking the two active sites. E72R mutation disrupts dimerization and abolishes activity.
  • Thermal robustness with fold melting temperature ~77 °C (nanoDSF), consistent with thermophilic origin.
  • Lid domain is minimal HBH; in the active structure the tunnel is open and accommodates C11–C18 of the MAG; in PMSF-inhibited structure, Tyr154 and Ser150 reposition to block the tunnel.
• Specificity determinants:
  • Glycerol binds via C2/C3 hydroxyls to Glu43, enforcing MAG selectivity (DAG/TAG sterically disfavored). Glu43 adopts an energetically strained conformation to mediate binding.
• Enzymatic activity data:
  • C8 MAG hydrolysis by WT: 16.4 ± 1.3 U mg−1; negligible activity on C8 DAG/TAG and only residual on C18:1 MAG in vitro (solubility-limited).
  • pNP acyl esters (C2–C18): Highest activities for C4–C8; residual for C16/C18. C8 pNP activity ~15% of C8 MAG activity.
  • Kinetics (pNP): KM decreases with chain length (from >2 mM for C2 to ~0.1 mM for C12–C14). kcat mirrors activity trend. Catalytic efficiencies for C4–C12 are similar (12–21 s−1 mM−1), with reduced efficiency for C2 and >C12, indicating solubility limits for longer chains.
• Mutational analysis:
  • S113A (active-site Ser) abolishes activity.
  • E43A retains ~1/3 of WT activity in vitro; E43K abolishes activity. In vivo, E43A shows ~2× higher FA release vs WT.
  • Y154A/R (lid tunnel) cause only minor effects on C8 MAG turnover in vitro; Y154R trends to increased FA release in vivo.
  • Lid deletion (Δ140–183) eliminates activity.
• Cellular assays:
  • Expression in E. coli increases total FA release (~2× vs S113A control). Major FAs detected: C16:0, C18:1; consistent with long-chain substrate turnover under cellular conditions.
• Comparative analysis:
  • Tth MAG lipase shares an HBH lid framework with cinnamoyl and feruloyl esterases but lacks their additional β-hairpin cap and polar residues critical for phenylpropanoid binding (Asp/Tyr), explaining different specificity. The Tth lid forms a tunnel enabling accommodation of long-chain MAGs, unlike the more occluded pockets of the phenylpropanoid esterases.

Capturing an endogenous long-chain MAG reaction intermediate in the crystal enabled definitive functional annotation of the orphan Tth enzyme as a long-chain monoacylglycerol lipase. The minimal HBH lid forms a hydrophobic tunnel that positions the distal acyl chain while the active site engages the glycerol backbone via Glu43, explaining strict preference for MAGs over DAG/TAG. Structural comparisons between active and PMSF-inhibited states show lid conformational plasticity that gates the tunnel. Mutational data validate key roles of the catalytic Ser113, dimerization (E72), and glycerol-binding Glu43 in catalysis and specificity. Kinetic measurements reveal increasing binding affinity with chain length but practical limits in catalytic efficiency for long-chain substrates due to solubility in vitro, whereas cellular assays support efficient turnover of C16/C18 acyl species. The work redefines a subclass of lipases/esterases with HBH lids, distinguishing them from canonical MAG lipases and phenylpropanoid esterases, and underscores how lid architecture and dynamics, beyond the catalytic triad, determine substrate routing and specificity. These insights are relevant for engineering lipases for biotechnological applications requiring robust activity and tailored substrate scope.

This study demonstrates that high-resolution structural biology can capture endogenous lipid intermediates to uncover true physiological function. The Tth enzyme is established as a non-canonical, long-chain MAG lipase featuring a minimal HBH lid whose hydrophobic tunnel guides long acyl chains to the active site. The structure–function framework, supported by mutagenesis, kinetics, and cellular assays, defines a prototype for an expanding family of HBH lid lipases/esterases and differentiates them from phenylpropanoid esterases. Practical implications include the ability to modulate activity via targeted mutations at the glycerol-binding site and lid tunnel, offering routes to design stable lipases for industrial processes. Future work should explore additional HBH family members, investigate membrane-proximal substrate acquisition, overcome solubility limits to quantify long-chain MAG kinetics, and harness lid engineering for expanded substrate promiscuity and improved turnover.

• In vitro assays for long-chain substrates (C16–C18) are limited by poor aqueous solubility, reducing measurable activity and complicating kinetic quantification.
• The endogenous ligand originates from the E. coli expression host rather than the native Tth environment; substrate profiles in the native organism were not tested.
• Structural inference of a tetrahedral intermediate is based on electron density and geometry without direct time-resolved validation; absence of the conserved catalytic water suggests possible reverse esterification during LC–MS preparation.
• Dimerization relevance in vivo is inferred from structural and SEC data; functional consequences in the native cellular context remain to be fully established.